See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/285234922

Simulation of a Micro-Scale Out-of-plane Compliant Mechanism

Conference Paper · October 2014

CITATIONS READS
10 408

5 authors, including:

Armando Arpys Arevalo Carreno Ehab Rawashdeh

ONI (Oxford Nanoimaging) King Abdullah University of Science and Technology
54 PUBLICATIONS   760 CITATIONS    8 PUBLICATIONS   85 CITATIONS   

SEE PROFILE SEE PROFILE

Nikolai Dechev
University of Victoria
40 PUBLICATIONS   1,107 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ultrasound Tracking of Internal Tissue Movement View project

Piezoelectric Ultrasonic Transducer Array View project

All content following this page was uploaded by Armando Arpys Arevalo Carreno on 30 November 2015.

The user has requested enhancement of the downloaded file.

Simulation of a Micro-Scale Out-of-plane Compliant Mechanism
David Castro1, Arpys Arevalo*1, Ehab Rawashdeh1, Nikolai Dechev2 and Ian G. Foulds1,3
1
King Abdullah University of Science and Technology,
Computer, Electrical and Mathematical Sciences and Engineering Division (CEMSE)
2
University of Victoria, Department of Mechanical Engineering.
3
The University of British Columbia, School of Engineering, Okanagan Campus
*Corresponding author: Thuwal 23955-6900, Kingdom of Saudi Arabia, arpys.arevalo@kaust.edu.sa

Abstract: In this work we present the simulation A Tsang suspension consists of a few
of a micro-scale large displacement compliant elements, which are: anchor pads, two
mechanism called the Tsang suspension. It symmetrical spring beams, and a central platform
consists of a flat micro-plate anchored down by (rectangular plate) as shown in Figure 1.
two springs on either side, that can rotate out-of-
plane and maintain its vertical assembly by a
simple single-axis actuation. COMSOL was used
to simulate these devices and extract the reaction
forces of the springs throughout the process of an
assembly. The effects of different design
parameters were investigated to help designers
understand the mechanical performance of such
a structure.

Keywords: MEMS, SU-8, compliant

mechanism.

1. Introduction

This paper investigates the effects of design

parameter variation on the mechanical
performance of the Tsang suspension [1,2].
Tsang suspensions are large displacement
compliant mechanisms that can be assembled
into out-of-plane configurations and provide a
platform that can be used for various purposes.
Tsang structures can be used in applications such
as micro-mirrors [3], free-space optics [4-6] and Figure 1. Top: SEM of an assembled SU-8 Tsang
RF systems [7]. Out-of-plane electro thermal Suspension. Bottom: Illustration of Tsang
suspension layout with Spring Length (SL) = 200 µm,
actuators have been fabricated using the Tsang
Spring Width (SW) = 30 µm, Number of Spring
suspension, where an actuator design was Beams (SB) = 4.
connected to the springs instead of the plate [8].
Tsang suspensions have also been used in Each spring beam is attached to the anchor
thermal isolation of sensors [9,10]. pad at one end, and the unanchored central
platform at the other end. The main advantage of
Similar out-of-plane structures, such as the this configuration is that it facilitates quick and
so-called buckled cantilevers have been used to easy assembly to achieve the final out-of-plane
provide similar advantages for sensors and position. This is achieved by single-axis
transducers [11-14]. However, the buckled actuation via an external micromanipulator, by
cantilever needs a secondary structure such as an pushing with a probe tip at the bottom of the
4@6;AD87
EFABB8D
FA
?4<@F4<@
<FE
V
AD<8@F4F<A@
plate, which will produce an out-of-plane
to the substrate, while the Tsang suspensions rotation into the final stable position.
create their own auto-locking mechanism due to
the reaction force generated by the springs once
the structure is assembled.

Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge

 Tsang suspensions are favorable for use,
since they are easily fabricated using a single
layer of micro-machined material. Further, they
can be easily assembled using automated
techniques, as described later in this paper. They
have been fabricated out of different materials
for various out-of-plane sensors and
microsystems [3-8,11]. This includes micro-
fabrication using SU-8 [11,13,14], poly (methyl
methacrylate) (PMMA) [12], polysilicon [1,15],
PMGI [16] and Polyimide [17-20].

To date, the design of previous Tsang

suspensions has been ad-hoc, and no work has
been presented to characterize and better
understand how design parameters affect the
mechanical performance of these structures.
Therefore, the work described here was
performed to characterize the design parameters Figure 2. Representation of the parameters that will
through simulation to help in choosing the be varied.
optimal parameters for the required application.

2. Design Parameters In order to facilitate comparisons between

The Tsang suspension is composed of
symmetric springs, an unanchored platform, and
the anchor pads (substrate), as shown in Figure
1. An in-plane force applied to the bottom edge
of the central platform produces a complex
deformation of the springs, which produces the
desired out-of-plane motion of the platform. As a
compliant mechanism, the geometry and
dimensions of these springs will have a great
impact on the structure’s performance, and will
therefore be the focus of this study.
The design parameters investigated in this
work were: spring length (SL), spring width
(SW), and the number of spring beams (SB), as
shown in Figure 2. For short the notation {SL,
SW, SB} will be used throughout the paper to
refer to a specific design. For example, the
notation {300, 20, 6} refers to a Tsang
suspension with SL = 300 µm, SW = 20 µm, and
SB = 6.
By varying SL, SW, and SB, the magnitudes
of the reaction forces produced by the springs in
the out-of-plane (z- direction) and in plane (y-
direction) were studied. Also, the lateral
displacement required to achieve a full out-of-
plane rotation was studied.
various designs, a “standard design” Tsang
suspension was established with the parameters
SL = 200 µm, SW = 30 µm, and SB = 4 {200,
30, 4}. This was used as the basis point for the
various parameter variations investigated. The
standard design was chosen since experience
with SU-8 has previously shown it as a reliable
and stable design. Keeping two parameters fixed
while changing the third, 10 different Tsang
suspension designs were investigated, with
variations of SL from 100 to 500 µm, SW from
20 to 50 µm and SB from 4 to 8, as shown in
Table 1. Tsang suspensions with an odd number
of spring beams have a fundamentally different
behavior and therefore, were not studied in this
work.
3. Use of COMSOL Multiphysics
One of the idiosyncrasies of Micro
Electromechanical Systems (MEMS) is the fact
that since device dimensions are small (typically
< 1 mm), direct measurement of their mechanical
properties can be challenging [21, 22].
Furthermore, this work deals with a large-
displacement compliant mechanism with
torsions, which can be quite complex to model
analytically, and therefore a common solution is

Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge

to resort to nonlinear finite element modeling
[23, 24].
The purpose of our simulations was to
determine the stress within the Tsang structure,
the reaction forces, and the displacement
throughout its assembly process, which aren’t
readily accessible through experimentation or
direct measurement alone. 
The simulation was
conducted with Finite Element Analysis (FEA)
using the COMSOL Multiphysics® software
package. For simulation, the Tsang structures
were first modeled using 3D computer aided
design (CAD) software, and those designs were
imported into COMSOL. The appropriate
boundary conditions that represented the
assembly process were applied, and the material
properties were specified. The anchor pads were
set such that the nodes in contact with the
substrate were fixed (i.e. 0 displacement at those
nodes in the x, y and z directions). The nodes
representing the lower edge of the platform were
constrained to allow for motion in the horizontal
direction, but no motion in the vertical direction.
This served to simulate the fact that the lower
edge of the platform remains in contact with the
substrate at all times. The actuation of the
micromanipulator was simulated by applying a
displacement boundary condition on the lower
edge of the platform in the horizontal direction.
The model was meshed with the automatic
tetrahedron mesh tool, and analysis was done
using the SPOOLES solver. Due to the large
displacements involved with Tsang structures,
the solver options for: (i) highly nonlinear and
(ii) large deformation, were selected in the
software. Figure 4 shows the simulation of an
assembled Tsang suspension.
Figure 4.Tsang suspension assembly in COMSOL.
To validate the simulation’s accuracy,
matching structures were fabricated, and a
Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge
Scanning Electron Microscope (SEM) was used
to capture top-view images of the assembled
structures. Figure 5 shows both an SEM image
and a simulation screenshot of an assembled
Tsang from the top. From this top-view image,
the angle of selected spring beams was measured
relative to a line drawn between the two anchor
pads. This angle is denoted as θa and θb, for the
simulation and SEM image, respectively. The
comparison revealed an average difference
between simulation θa and the experimental θb
angle of 1.12% with a standard deviation of
0.75%. This provided confidence that the
simulation adequately modeled the physical
device.
Figure 5. Top view of simulation and SEM image. An
example comparison angle, “θa” and “θb”, and
displacement to vertical “d” are shown.
4. Simulations Results
The simulations were used to estimate the
reaction forces of the Tsang structures, and
further, to estimate the effect of parameter
variation when designing the Tsang structures.
During the assembly process, the direction of the
spring reaction force changes as the rotation
angle of the plate increases. This reaction force
initially attempts to restore the plate to its
original flat position (backwards). However, a
critical “toggle point‟ (change-over point) is
reached, where with further rotation, the reaction
force begins to act downwards, thereby securing
the plate in an out-of-plane orientation. The first
step in the analysis of the Tsang structure was to
determine the angle of rotation of the platform
vs. the displacement of the lower edge of the
 Table 1: Design outputs for all 10 of the studied variations.
Lateral
displacement Ratio of Minimum
required for Max RFy Max RFz Max RFz to friction
Design assembly (µm) (mN) (mN) Max RFy coefficient
{100,30,4} 105 1.3 1.39 1.06 0.32
{200,30,4} 148 0.69 0.74 1.08 0.31
{300,30,4} 190 0.39 0.42 1.07 0.32
{400,30,4} 240 0.25 0.26 1.07 0.30
{500,30,4} 285 0.17 0.18 1.06 0.32
{200,30,6} 162 0.39 0.41 1.04 0.28
{200,30,8} 175 0.24 0.24 1.01 0.35
{200,20,4} 130 0.54 0.55 1.02 0.34
{200,40,4} 160 0.81 0.91 1.12 0.26
{200,50,4} 170 0.89 1.02 1.15 0.30
Thickness
variation*
5 µm 142 0.97 0.11 1.12 0.26
15 µm 160 1.95 1.93 0.99 0.28
20 µm 175 3.76 3.47 0.92 0.35
* Represents a variation from the standard design {200, 30, 4}.

platform. The results of one sample analysis for were not verified experimentally allowed
the design {200, 30, 4} are shown in Figure 6(a). exploring variations in the thickness of the
Figure 6(b) shows the reaction forces acting on springs. The results of these simulations are
the lower edge of the plate for the same design, included and shown in Table 1.
where Z denotes the vertical reaction force
(RFz), and Y the horizontal reaction force (RFy).
It can be seen that RFy increases to a maximum
then decreases, while RFz increases throughout
the assembly, reaching its maximum value when
the plate has reached its final position. The
decrease in RFy and increase in RFz is a
desirable feature, as it shows the ability of the
Tsang suspension to sustain its own assembly
configuration. As reported previously [1], a
measure of the structure’s stability can be
determined by the minimum coefficient of
friction required to sustain assembly, given by
the forces at the end of the rotation:

𝑅𝑅𝑅𝑅
,��˚
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚  𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓  𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ≥    (1)
𝑅𝑅𝑅𝑅
,��˚

A similar parameter is the ratio between the

maximum RFy and RFz reaction forces. The
maximum RFy is the force required to assemble
Figure 6. a) Angle of rotation versus lateral
the device, while the maximum RFz maintains
displacement of the Tsang suspension with SL = 200
assembly. The ratio between these two forces is a
µm, SW = 30 µm, SB = 4. b) Spring reaction force
measure of the conversion rate of the force versus lateral displacement, with same design
applied during the assembly, into a desirable parameters.
stabilizing force. Additional simulations that

Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge

 Figure 7. Graphs showing the effect of varying the different parameters as percentage variation of the standard
design {200,30,4}.
In Figure 7, a graphical summary of the
various design outcomes versus the change in
parameters SL, SW and SB is provided. The
design outcomes are: (a) Lateral displacement
required for assembly, (b) Maximum y force, (c)
Maximum z force, and (d) ratio of maximum z/y
force. The horizontal axis of all plots is listed as
a percentage, which is normalized with respect to
the standard design {200, 30, 4}. For example,
for variations in parameter SL, a value of 100%
would represent SL = 200 since the SL of the
standard design is 200. A value of 200 % would
represent a design with SL = 400, and so forth.
Each plot illustrates the effect of changing the
three design parameters, with respect to a
particular design outcome. For example, consider
the effect of increasing the spring-beam length.
Looking at the trends, it can be seen that
increasing the spring-beam length will lead to an
increase in the displacement required for
assembly, as shown in Fig. 7(a). It will also lead
to a decrease in the reaction forces in both z and
y directions, as shown in Fig. 7(b) and (c), and
has little effect on the Max RFz / Max RFy ratio,
shown in Fig 7(d). As another example,
increasing spring-beam width will result in an
increase in all parameters.
Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge
5. Conclusions
The Tsang suspension and its design
parameters were studied using COMSOL
Multiphysics. By varying the spring length (SL),
the spring width (SW) and the number of spring
beams (SB), changes in the reaction forces and
displacement required for assembly were
determined. From our simulation results, we
were able to generate graphs that can be used by
designers unfamiliar with Tsang structures, to
better understand the effect of changing a single
parameter. The general trends observed were as
follows. The in-plane reaction force of the
springs (RFy) increases rapidly at the start of the
assembly, reaches a maximum and gradually
decreases. The out-of-plane reaction force (RFz)
gradually increases throughout the assembly and
reaches its maximum value when the platform
approaches its vertical position. This change in
the forces is desirable as it produces a self-
sustaining vertical position on the platform
structure, and keeps it in a stable position. An
increase in any of the three parameters studied
(spring length, spring width, and number of
spring beams) will increase the structure’s
required lateral displacement, whereas reaction
forces, RFy and RFz, increase if spring width
increases, but decrease with the other
parameters. Varying these parameters shows
 only moderate effect on the structures stability as
measured by the minimum required coefficient
of friction. This work provides greater insight
into their operation and provides designers with
tools for designing their own implementation of
the Tsang suspension.
6. References
1. S. H. Tsang, D. Sameoto, I. G. Foulds, R. W.
Johnstone, and M. Parameswaran, “Automated
assembly of hingeless 90 degrees out-of-plane
microstructures,” Journal of Micromechanics and
Microengineering, vol. 17, no. 7, pp. 1314-1325.
(2007)
2. Tsang, S. H., D. Sameoto, I. G, Foulds, A. M.
Leung, and M. Parameswaran. "Wirebonder
assembly of hingeless 90◦ out-of-plane
microstructures." In Solid-State Sensors, Actuators,
and Microsystems Workshop, Hilton Head Island,
South Carolina, pp. 344-47. (2006)
3. S. Oak, G. F. Edmiston, G. Sivakumar, and T.
Dallas, "Rotating Out-of-Plane Micromirror,"
Journal of Microelectromechanical Systems,
vol.19, no.3, pp.632-639. (2010)
4. L. Y. Lin et al., “Micromachined integrated
optics for free-space interconnections,” Proc. IEEE
Micro Electro Mechanical Systems (MEMS ‟95)
pp 77–82. (1995)
5. M. C. Wu, “Micromachining for optical and
optoelectronic systems,” Proc. IEEE 85 1833–56.
(1997)
6. L. Y. Lin et al., “Free-space micromachined
optical switches with submillisecond switching
time for large-scale optical crossconnects” IEEE
Photon. Technol. Lett. 10 525–7. (1998)
7. D. J. Young et al., “Monolithic high-performance
three- dimensional coil inductors for wireless
communication applications” Technical Digest, Int.
Electron Devices Meeting, pp 67–70. (1997)
8. S. H. Tsang et al., "Out-of-plane electrothermal
actuators in silicon-on-insulator technology,"
Canadian Journal of Electrical and Computer
Engineering, vol.31, no.2, pp.97-103. (2006)
9. S. H. Tsang et al., "Thin Film Transistor (TFT)
Sensing Elements Fabricated in Surface
Micromachined Polymermems for a Differential
Calorimetric Flow Sensor," IEEE 22nd
International Conference on Micro Electro
Mechanical Systems, 2009. MEMS 2009.
pp.583-586, 25-29 Jan. (2009)
10. S. H. Tsang et al., "Monolithically fabricated
polymermems 3-axis thermal accelerometers
designed for automated wirebonder
assembly,"IEEE 21st International Conference on
Excerpt from the Proceedings of the 2014 COMSOL Conference in Cambridge
View publication stats
Micro Electro Mechanical Systems, 2008. MEMS
2008. pp.880-883, 13-17 Jan. (2008)
11. I. G. Foulds et al., “Polydimethylglutarimide
(PMGI) as a sacrificial material for SU-8 surface-
micromachining,” Journal of Micromechanics and
Microengineering, vol. 18, no. 7. (2008)
12. R. W. Johnstone et al., “Self-sacrificial surface
micromachining using poly(methyl methacrylate),”
Journal of Micromechanics and Microengineering,
vol. 18, no. 11. (2008)
13. S. W. Lee et al., “Lithographic stress control for
the self-assembly of polymer MEMS structures,”
Journal of Micromechanics and Microengineering,
vol. 18, no. 8. (2008)
14. D. Sameoto et al., “Integrated Testing Of
Polymer Mems Material Properties,” in
International Solid State Sensors and Actuators
Workshop, Hilton Head Island, 2008, pp. 174-177.
(2008)
15. S. Oak, “Testing and Characterization of 360°
Rotating Out-of-Plane Micromirrors,” Electrical
Engineering, Texas Tech University, Lubbock.
(2009)
16. I. G. Foulds et al., “Polydimethylglutarimide
(PMGI) as a structural material for surface-
micromachining,” Journal of Micromechanics and
Microengineering. vol. 18, no. 4, pp. 045026 (8pp).
(2008)
17. L. Marnat, A. A. A. Carreno et al., "New
Movable Plate for Efficient Millimeter Wave
Vertical on-Chip Antenna," IEEE Transactions on
Antennas and Propagation, vol. 61, pp. 1608-1615,
April 2013. (2013)
18. A. Alfadhel, A. A. A. Carreno et al., "Three-
Axis Magnetic Field Induction Sensor Realized on
Buckled Cantilever Plate," Ieee Transactions on
Magnetics, vol. 49, pp. 4144- 4147. (2013)
19. A. A. A. Carreno et al., "µHeater on a Buckled
Cantilever Plate for Gas Sensor Applications",
COMSOL Conference, Milan, Italy. (2012)
20. A. A. A. Carreno et al., "Optimized Cantilever-
to-Anchor Configurations of Buckled Cantilever
Plate Structures for Transducer Applications",
COMSOL Conference, Milan, Italy. (2012)
21. T. Yi, and C. J. Kim, “Measurement of
mechanical properties for MEMS materials”,
Measurement Science and Technology, 10(8), 706.
(1999)
22. K. J. Hemker and W. N. Sharpe Jr, “Microscale
characterization of mechanical properties”, Annu.
Rev. Mater. Res., 37, 93-126. (2007)
23. Howell, Larry L,”Compliant mechanisms”,
John Wiley & Sons, (2001)
24. Howell, Larry L. et al., “Handbook of
compliant mechanisms”. John Wiley & Sons.
(2013)